For those who have spent their career in the language industry – and it’s usually life-long – it can be incredible to think that today, barely a couple of decades into the digital revolution, machine translation is not only the norm, but commonplace: most people (at least in the first world) carry around in their pocket powerful, tiny computer, which can translate speech or text to almost any language in a second.
The disruption this has caused to localization is already immeasurable, and software – both in translation and the admin that goes with it – continues to transform the landscape, although not beyond all recognition, nor beyond our powers of prediction. Like the rest of digital technology, we can say with confidence the astonishing power of software is only going to get better.
Still, sometimes, we can feel like dinosaurs in our field – almost obsolete, and waiting for the inevitable AI asteroid impact.
But how much more are these changes going to disrupt the industry? A lot of the transition to new, digital ways of working has already happened, and surprisingly, hasn’t been totally destructive of what we had before. The prediction that apps like Google Translate would put human interpreters out of business has proved incorrect: the idea that AI can ever take over completely is now back in question.
Far from being an easy code to crack as we once thought, languages are increasingly recognized as not only infinitely nuanced, but ever changing, in ways that only a genuine intelligence can understand and interpret. Until artificial intelligences are created which are indistinguishable from human brains in their ability to understand and innovate, ushering in a Terminator-like dystopian future, it’s questionable whether real people will ever be out of a job.
Much localization is basically a creative process: although computers, using natural language processing, machine learning, sentiment analysis, statistics, and informed guesswork, might be able to translate predictable text like year-end financials, Tolstoy might still defeat them.
And while machine translation is very impressive indeed in the right context – preferably with clearly typed, grammatical sentences, or eloquently spoken words on crisp audio formats – it still has difficulty with the rough and tumble of the everyday word. The off the cuff interview recorded in a busy airport on a fuzzy microphone; the handwritten note. Humans will have a monopoly here for some time.
The strength – and the real disruption – inherent in increased computing power is in its ability to take away the drudgery of backbreaking repetitive tasks, both in active translation and in project workflow. As we discussed in the context of streaming services like Netflix, proprietary software for sourcing, testing, training and managing resources, and moving documents and media about are real game changers. They improve our ways of working; computers are colleagues, not competitors.
With new opportunities being thrown up by the internet all the time, this is important: in a world where news now breaks globally and instantly, and people demand instant access to customer reviews on international products and services, being able to move localization services fast is ever more vital. Software and plugins which integrate linguists, project managers, and commonly used platforms like WordPress, mean services like transcreation are truly viable enterprises – enabling linguists to co-create tailored content, or translate, in multiple languages, within hours or even minutes.
It’s no coincidence the word most often used for the way digitization is changing things is “disruption”, not “destruction”: this is an exciting time for localization, because it is an evolution, not a mass extinction. We are not wiping the slate clean and starting again with shiny robots, but finding new ways of integrating machine and man – truly the best of both worlds.